Offering (IPO) on March 24, 2010 First chip company founded after 9/11 to IPO Fastest chip company to IPO – 5.5 years Layout RF/MS IC CAD, Modeling & Process RFIC, DSP & Systems
▪0.13um digital CMOS process ▪Consumer, Mobile, Cable, and Automotive applications ▪July 2006, our first P.O. for $75K from Hauppauge for PCTV 2.5W (200+ Discretes) 250mW Single-Chip (10X less power) Multi-Channel Experience
CMOS DTTV Tuner 2009 MxL241 First Integrated Tuner and QAM Demodulator IC for Cable TV 2010 March - IPO MxL261 First 2 x 100MHz, 8-Channel Digital Cable Front End 2015 April – Acquired Entropic Communications, Inc. MxL248, MxL691/692 Industry’s 1st family of single-channel integrated ATSC tuner-demodulator ICs 2017 April - Acquired Marvell’s G.hn Home Networking and Access Business MxL1105/08 Industry's Only Single-Chip RF Transceiver Solution (CMOS) for the Entire Frequency Spectrum of 5 to 44GHz 2019 MxL1600 Quad-RF Transceivers with Widest Bandwidth and Lowest Power for 5G 2021 MxL9364x and MxL9368x Industry’s 1st 5nm CMOS 800G PAM4 DSPs 2012 MxL681/683 Highest Performance Single-Chip Receiver for ISDB-T Set-Top Boxes 2016 MxL3710 World’s 1st multi-gigabit MoCA 2.5, single chip, coax home networking solution 2018 MxL9354x Industry’s 1st 400G PAM4 DSP with Integrated Driver 2020 2022 Single-chip Solution for 5G Open RAN Radio Units MxL561 Industry's 1st Full-Spectrum Capture™ Satellite TV Receivers May – Acquired Broadcom’s Wireless Infrastructure Backhaul Business May - Acquired Exar Corporation MxL155x and MxL165x 1st RF Transceiver Solution with 8 transmitters, 8 receivers, and 2 feedback receivers in a single device August – Acquired Intel’s Home Gateway Platform Division September – Acquired NanoSemi, Inc. GPY241 Industry’s 1st quad-port PHY optimized for 2.5G Corporate Milestones Technology Milestones 2023 World’s 1st 2.5G Ethernet Switch with 8 Integrated 2.5G PHYs Puma™ 8 DOCSIS® 4.0 Cable Modem and Gateway Solution 2024 Single-Chip Radio Solution for all 4G/5G Applications
consumption with double the data rates from the previous generation 8x100G TIA with best-in-class crosstalk performance for high density applications Shipments poised to intercept data center 800G/1.6T upgrade cycles Bringing Design Efficiencies to High-Speed Interconnect
Encode / 25us Decode Latency* Up to 5X Higher Throughput Up to 2.5X Increased Compression Up to 3X Better Data Reduction Up to 2.5X Lower Latency *Note: Latency based on 32K block size Note: All Performance comparison vs Panther II
Off-Load Panther SDK API Driver Tools Linux and FreeBSD Support Provided in Source Code Fallover for All Panther Operations Optimized for AMD and Intel CPU Tools for Fast Evaluation & Integration
Off-Load Device Common FPGA based Only Compress or/and Encryption Panther Storage Accelerator Compression Encryption Deduplication Data Protection Real-Time Validation
Direct CPU Core Compression vs Panther Storage Accelerator 100 = $5K 3,140 = $157K Software LZ4 Compression (1.25:1) & SHA160 Software Deflate Compression (4:1) & SHA160 Panther XP10 (5:1), AES-CBC, Padding, SHA 256, T10-DIF, RTV Note: Software doing compression & SHA 160; CPU Core is estimated at $50 per core 2.0GHz Intel Xeon CPU Cores Throughput (Gbps) Common Software Solution Needed Software Solution Unmanageable CPU Requirement Unmanageable Solution Cost Uses 1000X More Cores Cores Limiting Data Throughput Minimal Data Compression Compress to Core Cost too High Saves $1,000s on Core Cost Improve Data Reduction Free Encryption, Data Protection & Real-Time Validation Increased Data Reliability
Hardware Off-Load Panther III Clear Winner Over Software Up to 1000X CPU Usage Reduction Up to 190X Lower Latency Up to 40X Higher Throughput Compression in Panther Free Encryption, Deduplication, Data Protection & RTV *Note: Latency based on 32K block size Note: All Performance comparison are software vs Panther III
& Deduplication Before Compression solutions were 2:1 Typical Deduplication were 2:1 Formula: Data Reduction = Compression X Deduplication 4:1 = 2:1 X 2:1 Now With Panther III Structured Data XP10 is 5:1 MaxHash Deduplication 3:1 15:1 = 5:1 X 3:1 Unstructured Data GZIP is 4:1 MaxHash Deduplication 3:1 12:1 = 4:1 X 3:1 12X Increase Effective Storage Same Drives Larger Effective Storage 12X Longer NVMe Drives Life Less Data Being Written Equals Increase The DWPD Rating 12X Faster Read & Write Speed Less Data Being Read / Written Increase Drive Processing To Complete 12X Faster Remote Replication Reduced Data Provides Faster Network Transfers And Lower Replicated Storage Cost
Effective Capacity NVMe Data Storage 25¢ per GB Panther III Enabled NVMe Data Storage 96TB Effective Capacity 2¢ per GB 81TB NVMe Drives Customer AI Storage Requirement: Cost Effective, Scalable, AFA For Fast Data Access Note: NVMe 960GB drive at $240 12X Data Reduction $21K Saving Drives Scalability 2U, 4U, 8U Drives
Standard Platforms • Half Height Half Length (HHHL) • Production Ready Solutions • Designed For Volume Production • Eliminate Chip Design Delay • Faster Integration with SDK • Fully Tested And Certified Boards
increase in data center power demand Goldman Sachs / Article / May 14, 2024 Data Centers Could Consume up to 9% of U.S. Electricity Generation by 2030 EPRI / PRNewswire / May 29, 2024 AI, data center electricity demand could drive advanced nuclear investment: NERC head Jim Robb Utility Dive / Dive Brief / June 6, 2024
Off-Load Device Common FPGA based Only Compress or/and Encryption Panther Storage Accelerator Compression Encryption Deduplication Data Protection Real-Time Validation
Clocking Kernel Mode 32K Block Size Transforms Compression GZIP Encryption AES-XTS Single Deduplication Hash SHA256 MaxHash Can not be tested on all 3 options Data Protection Can not be tested on all 3 options Real-Time Validation Can not be tested on all 3 options Power Comparison Setup 37 Common Platform and Transformation
Security + Deduplication + Data Reduction + Data Protection + Real-time validation (RTV) Enterprise Grade Ready SDK Driver, API, Tools 200Gbps Throughput; Scalability 3.2Tbps 40us Encode / 25us Decode Latency Panther III Improvements Up to 5X Higher Throughput Up to 2.5X Increased Compression Up to 2.5X Lower Latency Up to 3X Better Data Reduction Panther III PCIe Board Production Released Production Ready Solutions in Industry Standard Form Factors ready for Volume Production with compliancy and Certification Both PCIe HHHL & OCP SFF Board Solution ready Today Higher Level of Data Reduction 12X Increase Effective Storage Same Drives Larger Effective Storage 12X Longer NVMe Drives Life Less Data Being Written Equals Increase The DWPD Rating 12X Faster Read & Write Speed Less Data Being Read / Written Increase Drive Processing To Complete 12X Faster Remote Replication Reduced Data Provides Faster Network Transfers And Lower Replicated Storage Cost Panther III vs Software Up to 1000X CPU Core Reduction Up to 190X Lower Latency Up to 40X Higher Throughput Panther III Storage Server Power 17.7X Lower Power than Software Direct Core 2.1X Lower Power than competitive off-load Saves KW per Hour Add Improved Compression, Data Protection & Real-Time Validation with Minimal Power Increase
itpresstour
cryptopeg
masaruogura
sh_fk2
knih
oracle4engineer
shinyasaita
sonic
hamukazu
cheng_wei_chen
ninomiya_ii
asei
aeonpeople
reverentgeek
axbom
jcasabona
tamaranovitovic
schlessera
jonyablonski
erikaheidi
thirion
.png){kind=avatar}
helenjbeal
nwiizo
scottboms
aleyda
